Propulsion Control of Bionic Robotic Fish Based on Deep Deterministic Policy Gradient Algorithm

Abstract

Robotic fish exhibit considerable potential for a wide range of applications. However, the limitation of battery size highlights the need to improve swimming efficiency. This paper develops a Deep Deterministic Policy Gradient (DDPG)-based control method that makes the stiffness of robotic fish can be adjusted dynamically. First, the mathematical model of the two-joint robotic fish is established. Then, the conventional Proportional-Integral-Derivative (PID) control system and the deep deterministic policy gradient-based control system are developed. In the end, the feasibility of the Deep Deterministic Policy Gradient-based approach was validated through simulation and experiments. The results indicate that the control method improved the system efficiency by approximately 9.77%, suggesting that the proposed method holds promise as a high-efficiency propulsion control approach for robotic fish.